Heating diagnostics with MHD waves R. Erdélyi & Y. Taroyan Robertus@sheffield.ac.uk SP 2 RC, Department of Applied Mathematics, The University of Sheffield (UK)
The solar corona 1860s coronium discovered 1902 coronium has lesser atomic weight than hydrogen (Mendeleev) 1930s spectral lines due to known elements at very high stages of ionisation (Grotrian, Edlén) 2-5 December 2008 University of Sheffield
The heating enigma Courtesy: Unknown Nice Person
Loop structures are building blocks of corona Yohkoh/SXT TRACE Hinode/EIS
Theory of loop oscillations Coronal loops Main modes: n=0 fast sausage ( B, ρ) fast kink (almost incompressible) (Alfvenic) torsional (incompressible) slow (acoustic) type (ρ, v) n=1 Torsional waves
Theory of loop oscillations Closed tubes Main modes: sausage B > 0, ρ > 0 Expected oscillation periods in coronal loops sausage modes: P = 0.1 5 s kink ρ 0 kink modes: P = 1.4 14 min (Alfvénic) torsional ρ = 0 slow (acoustic) type B > 0, ρ < 0 V θ V z Alfvénic modes: P = 1 s slow modes: P = 7 70 min Aschwanden 2003, Wang 2004, Erdélyi 2008
Energetic & magneto-seismological implications of MHD waves Extraction and transfer of energy over long distances Energy deposition through various processes (nonlinear wave conversion and MHD shock formation, resonant absorption, phase mixing, etc.) Important dynamic (spicules, explosive events, etc.) and energetic consequences (plasma heating and acceleration) Coronal waves provide information about the magnetic field, transport coefficients, heating function, etc. diagnostics of corona
Heating footprints predicted by various mechanisms MHD waves Magnetic reconnection Wave-particle resonant interactions
Traditional methods for temperature measurements Filter ratio analysis Background subtraction Isothermal approximation
Case study: SXT and SUMER observations of hot loop oscillations P=8.1±0.3min V=30±11km/s
1D loop modelling (with TR!) 1 2 ), ( ) ( ) (, ) ( 0, ) ( 2 2 5/ 2 + = + + = + + = + = + γ ρ κ ρ ρ ρ ρ ρ ρ P v E L H s T T s vg Pv s Ev s t E g s P v s t v s t B B Thermal conduction Heating rate Radiation Chromosphere
Hydrodynamic evolution of the loop Footpoint excitation by pulse
SUMER measurements Results of line profile synthesis FeXIX CaXV CaXIII FeXIX CaXV CaXIII EIT 195Å
Results of combined observations and forward modelling (from Taroyan, Erdélyi et al. ApJ 2007) Establish the nature of the observed hot loop oscillations (Cauchy problem unique solution) Standing acoustic waves set up by a footpoint microflare Reproduce the time-distance profile of the heating function (important for understanding the nature of the heating process) Determination of heating function: micro-flare 1.5x10 27 erg Why only hot loops? Where are the SWs in cooler EUV loops?
Quest: Compute synthesized Hinode/EIS observations (i.e. theory) Question: Can standing waves exist in cooler loops? Selected lines -FeX 190Å, FeXI 188.23Å, FeXII 195Å Imaging mode - 40 slot Spectroscopy mode 1 slit Loop location disc centre Orientation south-north
Standing waves (P_pulse=P_f) Propagating waves (P_pulse<P_f) (both types of waves are formed but not very well seen in imaging mode see Taroyan et al. 2005 for theory) 2-5 December 2008 University of Sheffield
Best seen in spectroscopy mode Standing waves Propagating waves FeXII 195Å, 1 FeXI 188.23Å, 1 FeX 190Å, 1 Oscillations are seen in all three lines; Oscillations are in phase for Doppler but NOT in phase in intensity because of heating/cooling
Quest: Reality check #1 Hinode/EIS observations (i.e. real data) There is nothing to prevent oscillations in EUV loops Oscillations are best seen in Doppler shift let s analyse Hinode data!
XRT and EIS observations
Doppler shift along the slit x x
Doppler shift and intensity time series averaged over 5 pixels Quarter period phase shift! 1st example of acoustic waves in EUV loop cooling
Properties of wave: Wavelet analysis 1.2mHz
Wavelet analysis (w. background cooling)
Wavelet analysis (cooling substructed): 1.2 mhz
Interpretation The XRT and EIS snapshots suggest that The oscillations seen by EIS in spectroscopy mode correspond to a footpoint region of a loop; The oscillations are preceded by a microflare near the footpoint. Intensity increase in lower temperature lines (Fe VIII) and decrease in higher temperature lines (Fe XII, FeXIII, Ca XIII) Quarter period phase shifts between the intensity and Doppler shift oscillations. Most likely interpretation: standing acoustic wave in an EUV loop
Quest: Reality check #2 Hinode/EIS observations (i.e. kink oscillation)
XRT and EIS observations Slit close to apex detected wave motion is likely to be transversal No EIS images were available
Doppler shift and intensity time series averaged over 5 pxs Doppler shift interpreted as example of kink waves in EUV loop NO cooling
Wavelet analysis: 3 mhz
Interpretation The XRT snapshots suggest that the oscillations seen by EIS correspond to an apex region of a loop crossed by the slit; Transverse motions should have a line-of-sight component Doppler and no intensity oscillations --> magnetoacoustic kink waves Magnetic field measurements using intensity ratios between different Fe lines: B~10 G
Summary so far Hinode/XRT and Hinode/EIS observations in sit-and-stare mode are carried out to study oscillations in active region loops. Small amplitude oscillations are seen in different lines and pixels along the slit. Doppler shift and intensity oscillations (1 mhz) are detected near loop footpoints and are interpreted as standing longitudinal acoustic type waves. 3 mhz oscillations in the Doppler shift are present at near apex regions and are most likely to be kink waves. These waves have small amplitudes and have different origins from previously studied examples of flare-triggered oscillations. The oscillations are used to measure the magnetic field. (Erdélyi & Taroyan A&A 2008)
Problem The diagnostic methods currently used in coronal seismology rely on the presence of coherent standing/propagating waves The presence of such waves is not guaranteed! New method Analysis of Doppler shift time series as a diagnostic tool (What we propose to do when we do not see waves and oscillations)
Simple loop model: linear ideal (M)HD 0 < s < L (Wave operator) v = s Heating(t,s) Heating(t,s) = Σ Dirac delta or finite duration random pulses Boundary conditions at s=0, s=l
Green s function (Wave operator) G(t,s;τ,ξ) = δ(t-τ) δ(s-ξ) Boundary conditions at s=0, s=l Solution uniquely determined by the Green s function, BCs and the random Heating(t,s)
Results of the analytical study The Fourier analysis shows that the power spectrum for the time series of velocity has peaks at the frequencies of the standing waves(!) during random heating The height and shape of the peaks depends on the spatio-temporal properties (e.g., length, duration, spatial distribution, etc.) of heating
Analogy with helioseismology Corona? Major problems: Damping Localisation Limited observation duration
Case study: Results of a numerical study
Line synthesis Ne VIII (~0.8 MK), Mg X (~1.2 MK) resonant lines Emissivities along the loop Doppler width of the line Projection of bulk velocity on the line of sight Line broadening function Total intensity Line profiles
Uniformly random heating
Footpoint random heating
Multi-threaded loops Power ω1 ω2 ω3 2ω1 2ω2 2ω3 Frequency Distribution of frequencies would allow to derive info about fine-structure of the system
What we propose to do when we do not see waves and oscillations (Taroyan et al. A&A 2007) The power spectra to be used to Estimate the energies involved in the random pulses Study the multi-thermal structure of the loops Determine average loop temperature Distinguish uniformly heated loops from loops heated at their footpoints The power spectrum analysis of the Doppler shift timeseries for coronal line profiles is a potentially powerful tool for the diagnostics of the solar coronal plasma